The cellular responses of plants to numerous extracellular stimuli are mediated by transient elevations in the concentration of cytosolic free calcium ([Ca2+]cyt). We have addressed the question of how cells can use this apparently ubiquitous system to initiate so many specific and appropriate end responses. We show that the pollutant gas ozone elicits a biphasic Ca2+ response in intact Arabidopsis plants and a subsequent increase in expression of the gene encoding the antioxidant defence enzyme glutathione-S-transferase (GST). The second of the two [Ca2+]cyt peaks, but not the first, could be eliminated either by pre-treatment of plants with lanthanum chloride, or by reducing the duration of ozone fumigation. Under these conditions, ozone-inducedGSTexpression was abolished. These data provide a functional dissection of the ozone Ca2+ signalling pathway and indicate that the second ozone-induced [Ca2+]cyt peak provides the necessary information to direct expression ofGST.
It is now recognised that Ca2+ ions acting as intracellular second messengers couple a wide range of extracellular stimuli to their characteristic intracellular responses ( Trewavas & Malhó 1997; Webb et al. 1996). The apparent ubiquity of Ca2+-based signalling systems raises the question of how the cell is able to distinguish one calcium-mobilising stimulus from another ( McAinsh & Hetherington 1998). It is likely that a number of factors, including the cell’s developmental history, other second messengers and the cellular localisation of the elements that process the Ca2+ signal all contribute to the control of specificity ( McAinsh & Hetherington 1998; Snedden & Fromm 1998). Another possibility is that the stimulus-induced increases in the concentration of cytosolic free calcium ([Ca2+]cyt) [calcium signals] might contain information themselves. Such information could be encrypted in the form of oscillations or other tightly controlled spatio-temporal patterns ( McAinsh & Hetherington 1998). There is certainly evidence from studies that used Ca2+ indicators and the Ca2+-sensitive photoprotein aequorin that plants are capable of generating highly complex increases in [Ca2+]cyt ( Trewavas & Malhó 1997; Webb et al. 1996). However, it has not yet been possible to identify which component(s) of the calcium signal is responsible for directing a particular section of the final response. In order to address this issue, we have investigated the involvement of increases in [Ca2+]cyt in the signal transduction pathway(s) coupling the perception of the pollutant gas ozone to the induction of an antioxidant cellular defence mechanism.
Although ozone is essential in the stratosphere as a protectant against solar ultraviolet radiation, in the troposphere ozone is a pollutant which has serious effects on plant health ( Sandermann 1996). The toxicity of the gas results from its strong oxidising capacity, which leads to the production of highly reactive free radicals, and these species are responsible for initiating subsequent cellular damage ( Pell et al. 1997). Plants possess inducible endogenous defence mechanisms which are believed to protect them against the harmful effects of ozone ( Sandermann et al. 1998; Sharma & Davis 1997). Here we show that the induction of a gene for the antioxidant defence enzyme glutathione-S-transferase (GST) by ozone requires an increase in [Ca2+]cyt. Fumigation with concentrations of ozone that induce an increase in GST expression stimulates a biphasic increase in [Ca2+]cyt. We found that, on its own, the first peak in [Ca2+]cyt is insufficient for the induction of this gene by ozone and that the second peak in [Ca2+]cyt is required for increased GST expression. These data provide a functional dissection of a stimulus-induced Ca2+ signal and have allowed us to identify the precise component that is responsible for directing a particular part of the final response.
Results and discussion
To investigate whether ozone stimulated increases in [Ca2+]cyt we used Arabidopsis plants that had been stably transformed with the apoaequorin gene. When this gene is expressed in the presence of its luminophore coelenterazine, the reconstituted calcium-dependent luminescent protein aequorin ( Knight & Knight 1995) provides a non-invasive method for the measurement of [Ca2+]cyt. This approach has been used successfully on a number of previous occasions ( Johnson et al. 1995; Knight et al. 1991; 1996). Fumigation of Arabidopsis seedlings with ozone caused increases in [Ca2+]cyt that were directly proportional to the concentration of the pollutant applied ( Figs 1 and 2). Exposure to 35 p.p.b. ozone induced a single short-lived spike-like elevation of [Ca2+]cyt ( Fig. 1b), whereas at ozone concentrations of 70 p.p.b. or higher the spike was followed by a second more gradual elevation which, although smaller in magnitude, was considerably longer in duration ( Fig. 1c–e). In contrast, fumigation of transformed seedlings with ozone-free air failed to generate any increase in [Ca2+]cyt ( Fig. 1a).
The concentrations of ozone used in these experiments caused no visible damage to the Arabidopsis seedlings (data not shown). Therefore, we were able to dismiss the possibility that the ozone-induced elevations of [Ca2+]cyt formed part of a cytotoxic response to this pollutant. This prompted us to investigate whether the elevation of [Ca2+]cyt formed part of an intracellular signal transduction pathway activated by ozone and which led to the induction of cellular defence mechanisms. To answer this question we took as our focus the gene for the major detoxification enzyme GST ( Marrs 1996; Noctor & Foyer 1998). Previous work had shown that fumigation with ozone resulted in an increase in GST mRNA transcript abundance in Arabidopsis and GST activity in barley ( Conklin & Last 1995; Price et al. 1990; Sharma & Davis 1994). Using reverse-transcriptase polymerase chain reaction (RT-PCR) we observed that ozone concentrations greater than 70 p.p.b. induced an increase in GST expression in the apoaequorin transformed Arabidopsis seedlings ( Fig. 3). In order to seek evidence that Ca2+ ions were acting as second messengers in the signal transduction pathway that links ozone perception with the induction of GST expression, we first investigated whether the Ca2+ agonist BayK8644 (which induces prolonged elevation of [Ca2+]cyt in Arabidopsis seedlings; C. Podmore, H. Knight and M.R. Knight, unpublished results), could mimic the actions of ozone. Figure 3 shows that when the seedlings were treated with BayK 8644 there was an increase in GST transcript abundance. In contrast, when the seedlings were treated with water as a control there was no detectable increase in GST expression. These data suggest that Ca2+ ions have the potential to act as second messengers in the induction of GST expression by ozone.
From the data in Figs 1 and 2 it is apparent that there is a correlation between the appearance of the second ozone-induced elevation of [Ca2+]cyt and the increase in GST expression. However, this correlation is not complete because in three out of five experiments a small but statistically significant second peak ( Fig. 2) was observed after the seedlings were exposed to 70 p.p.b. ozone. In contrast, fumigation with this concentration of ozone did not result in a detectable increase in GST transcripts. This apparent discrepancy can be interpreted in two different ways. Either the small [Ca2+]cyt increases elicited by 70 p.p.b. ozone are only able to induce a very low level of GST expression that may be too small to detect above basal levels ( Figs 3 and 4), or the increase in [Ca2+]cyt induced by exposure to 70 p.p.b. ozone is itself below the level required to trigger the transcription of the GST gene. This latter possibility implies the presence of a threshold of [Ca2+]cyt that needs to be breached before any activation of GST can occur.
In order to test directly whether the second elevation was required to trigger the increased GST mRNA levels, we employed two strategies that would allow us to set the experimental conditions such that the second elevation would not appear or would be greatly reduced even after exposure to 200 p.p.b. ozone. Figure 4(b) shows that when the plants were fumigated with 200 p.p.b. ozone in the presence of La3+ ions, the second peak was almost completely abolished although the first spike was still present. Under these conditions there was no detectable increase in GST expression ( Fig. 4d). In contrast, fumigation with 200 p.p.b. ozone in the absence of La3+ ions produced the expected biphasic [Ca2+]cyt response and induced an increase in GST expression ( Fig. 4a,d). In our second approach we took advantage of the fact that a brief (4 min) exposure to 200 p.p.b. ozone only causes the appearance of the primary ozone-induced spike of [Ca2+]cyt ( Fig. 4c). When GST transcripts were quantified under these experimental conditions, it was apparent that there was no increase in GST expression ( Fig. 4e). Taken together, these data provide strong evidence that the second ozone-induced peak is required for the induction by ozone of GST expression and that the first peak in isolation does not contain the signalling information required to direct the expression of this gene.
The focus of this work has been the question of how specificity is encoded in Ca2+-based signalling systems ( McAinsh & Hetherington 1998). Current studies have shown that ozone induces, in temporal terms, a biphasic calcium signal. Interestingly, Sedbrook et al. (1996), using the recombinant aequorin approach, observed a biphasic calcium signal in response to anoxia. However, in contrast to our work Sedbrook et al. (1996) found that changes in [Ca2+]cyt did not correlate strictly with the activation of gene expression.
We conclude that the second peak of the ozone-induced increase in [Ca2+]cyt stimulates GST expression. This raises the important question: How does the plant discriminate between these two peaks of [Ca2+]cyt? One mechanism would be that each peak is localised to a different cell type. In this study, the expression of the calcium reporting apoaequorin gene is constitutive and, therefore, it is not possible to determine whether the Ca2+ signal that we have measured is common to all cells or represents the summation of different Ca2+ signals specific to individual cell types. Consequently, the first and second peaks might be associated with different types of cell but such localisation is masked in the system we used. In order to investigate this possibility, it will be necessary to either conduct detailed imaging on the existing constitutively expressing plants or to use plants expressing the apoaequorin gene under the control of cell-specific promoters. In this latter context, it will also be important to investigate ozone-induced GST expression at the level of individual cell types and to determine whether the correlation between expression and presence of the second peak exists at the cell-type level. Another possible explanation for the biphasic calcium signal is that as the ozone diffuses through the leaf it will trigger calcium elevations in the different cell layers at different times that correspond to when individual tissues perceive the gas. However, we feel that if this were the case it would be more likely to generate a sustained increase in Ca2+ rather than a series of discrete elevations.
If the biphasic signal is present in single cells there are two mechanisms that might allow a cell to distinguish between the spike-like first elevation and the prolonged second increase in [Ca2+]cyt observed in response to ozone. One possibility is that the spike-like elevation may not be of sufficient duration or magnitude to set in motion the train of events that culminate in increased transcription. Evidence to support such a possibility has been obtained from guard cells. When these cells were injected with water they exhibited a transient spike-like increase in [Ca2+]cyt but did not reduce cell turgor. In contrast, microinjection with cyclic-ADP ribose induced a sustained elevation in [Ca2+]cyt which was associated with closure of the stomatal pore ( Leckie et al. 1998). Another possibility is that different parts of the Ca2+ signal are localised to different subcellular regions. There are data from animal cells that demonstrate the importance of localised increases in calcium in dictating the final cellular response. For example, in mouse AtT20 cells it has been shown that elevations in nuclear Ca2+ control Ca2+-activated gene expression via the cyclic AMP response element, whilst increases in [Ca2+]cyt regulate gene expression through the serum response element ( Hardingham et al. 1997). Recently it has also been demonstrated that the amplitude and duration of Ca2+ signals differentially control the activation of transcriptional regulators ( Dolmetsch et al. 1997; 1998). The extent to which either of these possibilities relates to ozone signal transduction must await the results of future experiments.
In this series of experiments we have shown that ozone induces a biphasic calcium signal and that successful coupling of ozone perception to GST expression requires the presence of the second ozone-induced elevation in [Ca2+]cyt. This represents a functional dissection of a stress-induced calcium signal in plants and suggests that temporal elements in the Ca2+ signal are important in dictating the outcome of the final response.
Growth and fumigation
Seedlings of Arabidopsis thaliana transformed to express apoaequorin in the cytosol ( Knight et al. 1991 ; Johnson et al. 1995 ; Knight et al. 1996 ) were grown on plates of full-strength Murashige and Skoog medium (Sigma), 0.8% agar for 10 or 14 days then pre-incubated in 2.5 μm coelenterazine overnight to reconstitute the Ca2+-active photoprotein aequorin ( Knight et al. 1996 ). Individual seedlings were fumigated in a 3.5 ml cuvette. The ozone-enriched air was bubbled through water to remove any hydrogen peroxide produced by the generator. Luminescence was measured using a digital chemiluminometer with a photomultiplier ( Knight et al. 1996 ). A steady background luminescence was recorded for 2 min before introducing the ozone. The flow rate was 100 ml min–1 throughout the hour. At the end, 1 ml of 900 m m CaCl2 10% ethanol was added to the sample cuvette to discharge any remaining aequorin. Calibrations were performed using a calibration equation derived empirically ( Knight & Knight 1995; Knight et al. 1996 ). In order to control against the possibility of direct calcium-independent discharge of the coelenterazine provoked by ozone ( Lucas & Solano 1992), wild-type plants were treated with coelenterazine and fumigated with ozone. These plants produced no detectable luminescence response to ozone (data not shown). Additionally, fumigating a plant homogenate containing reconstituted aequorin tested the possibility that ozone was provoking a direct calcium-independent discharge of the aequorin. This treatment did not produce detectable discharge of the aequorin (data not shown). Finally, the possibility that hydrogen peroxide contamination was the cause of the [Ca2+]cyt response to fumigation was tested by passing the already water-scrubbed ozone through 1 ml of water for 1 h to collect any potential hydrogen peroxide. The addition of this water to reconstituted Arabidopsis seedlings produced no luminescence response (data not shown).
Analysis of transcript abundance using inverse PCR
Seedlings were fumigated in batches of 20 in 5 cm diameter dishes at either 35 or 200 p.p.b. ozone, 400 ml min–1 for 58 min in the dark. The controls were fumigated for 1 h with ozone-free air. The higher flow rate was used to compensate for the larger volume under fumigation. Total RNA was extracted 3 h after the end of fumigation using a Qiagen RNeasy Plant total RNA extraction kit as recommended by the manufacturer. cDNA was synthesised from 1 μg total RNA from each sample as described previously ( Knight et al. 1996 ). This cDNA was serially diluted and 1:1000 dilutions were subjected to RT-PCR using primers for GST and 1:10 000 dilutions using apoaequorin-specific primers, AEQ (as an internal standard) as described previously ( Knight et al. 1996 ). Both these dilutions were found to produce an exponential amplification with the respective primers. The GST-specific primers used were 5′TTGCTTCTTGCTCTTAACCC3′ and 5′CTCAACCTTCTCCAAATTCC3′. Amplified DNA products were separated on a 1% (w/v) agarose TBE gel.
A.M.H., M.R.M. and H.C. are grateful to the UK Natural Environment Research Council for the award of a research grant. H.K. and M.R.K. are grateful to the BBSRC for the award of a research grant. M.R.M. and M.R.K. gratefully acknowledge the Royal Society of London for the award of University Research Fellowships.